Yazar "Baykan, Nurdan Akhan" seçeneğine göre listele
Listeleniyor 1 - 11 / 11
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Çapraz ilişki metoduyla iris tanıma(Selçuk Üniversitesi Teknik Bilimler Meslek Yüksekokulu, 2011) Yıldız, Ferruh; Baykan, Nurdan AkhanBu çalışmada, biyometrik sistemlerden iris tanıma için yeni bir algoritma geliştirilmiştir. Tanıma işlemleri için kullanılan göz resimleri “National Laboratory of Pattern Recognition - Institute of Automation Chinese Academy of Sciences” laboratuarının CASIA Iris veritabanından alınmış olup, 320 x 280 piksel boyutlarında gri seviye resimlerdir. Bu çalışmanın uygulamasında 50 farklı kişiye ait, toplam 150 resimden oluşan veri tabanı üzerinde çalışılmıştır. Kullanılan göz resimlerinden, tanıma işleminde kullanılacak olan iris kesit alanlarını tespit etmek için yeni bir segmentasyon algoritması geliştirilmiş ve böylece tüm bir göz resmi içinden, sadece iris alanına ait 64 x 64 piksel boyutlarında kesit alanlar elde edilmiştir. Böylece tanıma işlemi için tüm bir resim değil, sadece irise ait kesit alanlar kullanıldığından, hem işlem süresi kısaltılmış, hem de tanıma işlemi esnasında test edilen piksel sayısı için yapılacak işlem miktarı azaltıldığı için performans arttırılmıştır. Bu kesit alanlara göre karşılaştırma ve tanıma yapabilmek amacıyla, resimdeki piksel değerlerini kullanarak işlem yapan Çapraz İlişki (cross correlation) metodu ve pikseller arasındaki hataların kareleri toplamının ortalama karekökü hesaplamaları kullanılmıştır. Çapraz İlişki metodunda korelasyonun en büyük ve ortalama hatanın da en küçük olduğu görüntülerin doğru görüntü olarak alındığı programda, uygulama sonucunda, aynı kişiye ait, farklı zamanlarda çekilmiş, farklı göz resimlerinde % 77 benzerlik oranına kadar doğru tanıma gerçekleştirilebilmiştir. Elde edilen sonuçlara göre, Çapraz İlişki modelinin kullanılacağı uygulamalarda, dijital görüntülerde dönüklüğün 200’den büyük, ölçek farklılığının ise %30’dan daha büyük olduğu resim koşullarında tanımanın zorlaştığı, hatalı sonuç alınabildiği gözlenmiştir.Öğe Case Study in Effects of Color Spaces for Mineral Identification(Academic Journals, 2010) Baykan, Nurdan Akhan; Yılmaz, Nihat; Kansun, GürselColor is the first parameter and one of the most powerful and important feature for mineral recognition via image processing. Although there are different color spaces, the most used of these are, three color spaces, namely RGB, HSV and CIELab were compared to find the best color space for the mineral identification in this study. These three color spaces are compared in terms of their suitability for identification. Using these three color space, an artificial neural network is used for the classification of minerals. Optical data of thin sections is acquired from the rotating polarizing microscope stage to classify 5 different minerals, namely, quartz, muscovite, biotite, chlorite, and opaque. The results show that RGB was efficient and suggested as the best color space for identification of minerals.Öğe Effects of Color Spaces and Distance Norms on Graph-Based Image Segmentation(IEEE, 2017) Saglam, Ali; Baykan, Nurdan AkhanUse of the graph theory tools in image processing field is growing up with each passing day. Graph theory makes the operations easier for image processing applications, and can represent digital image components completely. In image segmentation processes, the graph theory tools are also used widely. These kinds of image segmentation processes are called graph-based image segmentation. In many image processing applications, it seems as a problem that which color space the color values of pixels should be considered according to and which distance norm should be used to measure the difference between two points in the space. In this work, a graph-based image segmentation algorithm is tested on several color spaces with different distance norms. The test is carried out on 100 real world images that take part in a general-purposed image segmentation dataset. The average segmentation results are given as F-measure in this work with regard to both color spaces and distance norms. The results show that L*a*b* and L*u*v* color spaces are more appropriate than RGB color space, in general. The squared Euclidean distance norm gives more accurate results than the Euclidean distance norm, used in the source paper, if the Gaussian smoothing is not used as pre-processing.Öğe An Efficient Object Extraction with Graph-Based Image Segmentation(IEEE, 2015) Saglam, Ali; Baykan, Nurdan AkhanObject extraction process is a closely related issue with image segmentation process. To separate an image to several segments formed similar pixels, many methods are proposed in the area of image processing. Graph-based image segmentation is also one of the segmentation methods. Because of their representation convenience and ease of use, graphs are used as important tools in many image processing applications. While an image segmentation process runs, the processes splitting a graph to sub graphs and merging sub graphs are carried out in the meanwhile. To fulfill these processes, the method uses some local features such as differences between vertices in the graph, which represent pixels, or global features of the image and its segments. To extract an object from an image, we first segmented the entire image, because of to save global features, or to obtain more accurate segmentation. Finally, we extract the intended object from the image by merging the segments that are inside the area drawn before by us. We test the method on some images in the Segmentation Evaluating Database from Weizmann Institute of Science and evaluate the segmentation results. Our F-measure score values show that it seems noticeable good segmentation.Öğe Fast K-means color image clustering with normalized distance values(Selçuk Üniversitesi Mühendislik Fakültesi, 2018) Baykan, Nurdan Akhan; Sağlam, AliImage segmentation is an intermediate image processing stage in which the pixels of the image are grouped into clusters such that the data resulted from this stage is more meaningful for the next stage. Many clustering methods are used widely to segment the images. For this purpose, most clustering methods use the features of the image pixels. While some clustering method consider the local features of images by taking into account the neighborhood system of the pixels, some consider the global features of images. The algorithm of the K-means clustering method, that is easy to understand and simple to put into practice, performs by considering the global features of the entire image. In this algorithm, the number of cluster is given by users initially as an input value. For the segmentation process, if the distribution of the pixels over a histogram is used, the algorithm runs faster. The values in the histogram must be discrete in a certain range. In this paper, we use the Euclidean distance between the color values of the pixels and the mean color values of the entire image for taking advantage of the every color values of the pixels. To obtain a histogram that consists of discrete values, we normalize the distance value in a specific range and round the values to the nearest integers for discretization. We tested the versions of K-means with the gray-level histogram and the distance value histogram on an urban image dataset getting from ISPRS WG III/4 2D Semantic Labeling dataset. Comparing the two histograms, the distance value histogram proposed in this paper is better than the gray-level histogramÖğe A mineral classification system with multiple artifical neural network using k-fold cross validation(2011) Baykan, Nurdan Akhan; Yılmaz, NihatTürkçe Öz YokÖğe Mineral Identification Using Color Spaces and Artificial Neural Networks(PERGAMON-ELSEVIER SCIENCE LTD, 2010) Baykan, Nurdan Akhan; Yılmaz, NihatIdentification of minerals and percentage of their area within a thin section of rock are important for identifying and naming rocks. Colors of minerals are the basic factors for identification. In this study, an artificial neural network is used for the classification of minerals. Optical data of thin sections is acquired from the rotating polarizing microscope stage. For the first analysis we selected a set of parameters based on red, green. blue (RGB) and the second based on hue, saturation, value (HSV) color spaces are extracted from the segmented minerals within each data set. A neural network with k-fold cross validation is trained with manually classified mineral samples based on their pixel values. The most successful artificial network to date is the three-layer feed forward network which uses minimum square error correction. The network uses 6 distinct input parameters to classify 5 different minerals, namely, quartz, muscovite, biotite, chlorite, and opaque. Testing the network with previously unseen mineral samples yielded successful results as high as 81-98%.Öğe Multiclass Mineral Recognition Using Similarity Features and Ensembles of Pair-Wise Classifiers(SPRINGER-VERLAG BERLIN, 2010) Kybartas, Rimantas; Baykan, Nurdan Akhan; Yılmaz, Nihat; Raudys, SarunasMineral determination is a basis of the petrography Automatic mineral classification based on digital image, analysis is getting very popular To improve classification accuracy we consider similarity features. complex one stage classfiers and two-stage classifiers based on simple pair-wise classification algorithms Results show that employment of two-stage classifieis with proper parameters or K class single layer perceptron are good choices for mineral classification Similarity features with properly selected parameters allow obtaining non-lineal decision boundaries and lead to sizeable decrease in classification error rateÖğe Roof Detection on Satellite Images(IEEE, 2017) Saglam, Ali; Baykan, Nurdan AkhanIn digital image classification processes, image pixels are separated according to some features that they have. Satellite images are digital images that are taken by a satellite vehicle through some sensors which perceive the specific wavelength of the light. In this study, two different digital image classification method (Linear Discriminant Analysis and Normalized Distance Values) have compared to each other, using different color spaces (RGB, L*a*b* and HSV), on the satellite images that have been taken by digital airborne sensors so as to detect roof objects. The common features of the applied methods are those they are supervised because of using training data given previously and they run fast because of operating linearly using a threshold value. For this reason, some of the images in the dataset are used for the purpose of training in order to detect the certain coefficients and the threshold value. The dataset we used for training and testing are the images acquired from ISPRS WG III/4 2D Semantic Labeling database. In the database, the classification ground truth images are also available.Öğe Sequential image segmentation based on minimum spanning tree representation(ELSEVIER SCIENCE BV, 2017) Saglam, Ali; Baykan, Nurdan AkhanImage segmentation is a very important stage in various image processing applications. Segmentation of pixels of an image and clustering of data are closely related to each other. For many graph-based data-clustering methods and many graph-based image-segmentation methods, minimum spanning tree (MST)-based approaches play a crucial role because of their ease of operation and low computational complexity. In this paper, we improve a successful data-clustering algorithm that uses Prim's sequential representation of MST, for the purpose of image segmentation. The algorithm runs by scanning the complete MST structure of the entire image, such that it finds, and then cuts, inconsistent edges among a constantly changing juxtaposed edge string whose elements are obtained from the MST at a specific length. In our method, the length of the string not only determines the edges to compare, but also helps to remove the small, undesired cluster particles. We also develop a new predicate for the cutting criterion. The criterion takes into account several local and global features that differ from image to image. We test our algorithm on a database that consists of real images. The results show that the proposed method can compete with the most popular image segmentation algorithms in terms of low execution time. (C) 2016 Elsevier B.V. All rights reserved.Öğe User Interactive Object Extraction with Sequential Image Segmentation(IEEE, 2018) Saglam, Ali; Baykan, Nurdan AkhanIn this study, a graph-based image segmentation algorithm which was developed in recent years and achieves a significant success in terms of the performance of both the accuracy and speed is used as an intermediate process for a user interactive object extraction method. In the object extraction method developed, the related image is subdivided into segments and, then, these segments are merged according to their label values by using the area determined by the user at first. The image segmentation algorithm used in the scope of this work fulfills a sequential segmentation process on the one dimensional edge array of Prim's minimum spanning tree representation. The algorithm does the segmentation by cutting the specified edges on the tree. According to the method developed, these cut edges are kept and some of them are added to the tree again in the merging stage; so that, the segments at the ends of the edge added are merged. Owing to this process, the process of finding the least weighted edge between the two segments to be merged, which needs to be performed before the merging stage according to the previous studies, is not needed. The method developed here is compared with some methods in the literature on a dataset consist of real life images, and it seems that the method shows a significant superiority to the other methods.
